This page is based on "Petrophysics and Fracture Optimization", by D. Holgate,  P.Geol. and E. R. Crain, P.Eng. Half day seminar presented by Dorian Holgate at the SPE/CSUR Unconventional Resource Conference, Calgary, Alberta, Sept. 2014.

Completion success depends on accurate parameters determined from petrophysical analysis. Many stimulation designs are faulty because of poor quality sonic and density log data. Raw log data are often inadequate due to rough borehole conditions and light hydrocarbon effect.
 Stimulation design software expects data for the water filled case, so gas corrections may be needed.

In some unconventional reservoirs, the presence of kerogen confounds standard log analysis models. Kerogen looks a lot like porosity to most porosity-indicating logs. A single log, or any combination of them, will give highly optimistic porosity and free-gas or oil saturations unless appropriate corrections are applied.

This tutorial explains how to deal with poor quality sonic and density data, for the purpose of calculating mechanical rock properties, for input to hydraulic frac software modeling packages (GOHFER, FRACPRO and MFrac).

Complications tied to kerogen rich reservoirs are also examined. The shale and kerogen-corrected model is presented as a solution. When run in a logical step-by-step procedure, as opposed to a multi-mineral solution, repairs to input data and parameter adjustments become quite obvious and straight forward because core and lab data can be used at each point along the route. To accomplish this task reliably, we created what we call "Holgate and Crain’s 20 Step Petrophysical Workflow".

When the analysis is complete, we can start to design a hydraulic fracture that might actually behave as predicted. Note that if the calibrations at each step are performed, the need for re-runs will be quite low. The technique presumes, of course, that you did acquire and process all the needed lab data. If you do not have it all, you may survive. If you have none of it, you may as well wave your arms over some goat entrails.

Step 1: Data Gathering / Inventory
Logs required are a full suite with resistivity, compressional and shear sonic, density, neutron, nuclear magnetic, spectral gamma ray, and caliper logs. LAS (Log ASCII Standard) files must be reviewed for curve availability. A text editor (Notepad, Wordpad) can be used to open LAS files to review curve data and borehole parameters. Measured depth logs should always be loaded, along with a deviation survey allows reference between MD, TVD, and TVDSS.

Rock data required over the zones of interest comprise conventional and/or tight rock core analysis (porosity and permeability), XRD for mineralogy and clay mass fraction, geochemical analysis for TOC and Ro, and sample descriptions.

Key wells with required data should be chosen and attempts made to gather missing data.

A three well minimum is recommended for projects. Rarely will the subject well have all data needed to complete a calibrated petrophysical analysis. Offset wells should always be reviewed and used to put together the best data set possible. The accuracy of the petrophysical model improves with an increased number of wells reviewed.

Read More about Data Gathering  


Step 2: Quality Checks

Logs must be checked for depth control, an expanded-scale depth plot is very useful for this. Compatible porosity scales must be used (quartz, calcite, or dolomite). Units must be consistent for all logs (english or metric). Logs may need to be normalized. NULL values and spikes over short intervals need to be fixed.

Plot all available curves to check data integrity
        and to determine depth adjustments

Read More about Data Quality Control


Step 3: Identify Intervals with
      Questionable Data

Caliper and density correction logs are used to identify borehole intervals which are washed out (larger diameter than the drill bit). Calculation sequence may need to be modified over these intervals. Reconstructed logs are often required.



Shaded intervals indicate zones with suspect data that need attention

Read More about Editing Logs








Step 4: Calculate Volume
     OF Shale

Petrophysicists define volume of shale as the bulk volume of the rock composed of clay minerals and clay bound water. Gamma ray log is typically used to calculate shale volume. A non-linear relationship between shale and clean endpoints is required for radioactive intervals (Clavier, etc.).  A spectral gamma ray log is the most useful for determining shale volume over radioactive intervals. Thorium is associated with clay, potassium is associated with feldspar and clay. Uranium is associated with organics (TOC).

Volume of shale can also be calculated from the SP log, resistivity log, and separation between neutron and density logs. However, in organic-rich reservoirs, these methods are usually not useful. Thorium curve is best, followed by total gamma ray method.
Calibrate results to bulk clay from XRD analysis. If XRD has not been run, it can be done using rock chip samples or cores from subject or offset wells.  Adjust parameters or use alternate Vsh model  to obtain a better match.

Read More about Shake Volume

Montney interval displaying XRD calibrated
shale volume - black dots in right hand
 track are XRD bulk clay volume, converted
 from mass fraction data

Black dots in Track 4 show TOC from lab, brown shading is kerogen volume and red shading is free gas in pore space. Blue dots are core porosity.




Step 5: Calculate Kerogen Volume

Total organic carbon mass fraction can be calculated from the resistivity log and a porosity log, using Passey or Issler methods. The Passey model is often called the “DlogR” method, with the “D” standing for “Delta-T” or sonic travel time. Passey also published density and neutron log versions of the equations. Baseline log values are required and are supposed to be picked in non-source rock shales in the same geologic age as the reservoir. Unfortunately, baseline values are often not available, makes the Passey model difficult to calibrate. Level of organic maturity (LOM) is also required, but is seldom measured, except as vitrinite reflectance (Ro). There is a general relationship between LOM and Ro, so an estimate can be made, with a little trial and error. .LOM is in the range of 6 to 11 in gas shale and 11 to 18 in oil shale. A higher LOM will give a lower TOC when using the Passey method.
Issler’s method, which is based on WCSB Cretaceous data is preferred as no baselines are needed. Both Issler and Passey models may require a scale factor for older rocks.


Graphs illustrating Issler's TOC relationship

Bear in mind that lab TOC measures only the carbon content in the kerogen. Kerogen mass fraction is larger than TOC mass fraction due to other elements,
such as oxygen, nitrogen, sulphur, etc., in the kerogen component. The conversion factor is the ratio of carbon weight to total kerogen weight. Again this is not routinely measured in the lab; typical range is from 0.68 to 0.95, default 0.80. KERwt = TOCwt / Factor.

Kerogen mass fraction is then converted to volume fraction using a density in the range of 1200 to 1400 kg/m3.

TOC mass fraction results to geochem TOC analysis. Adjust parameters or use scale factor to obtain a better match.

Read More about TOC and Kerogen .

Montney interval displaying calibrated TOC weight fraction and the associated kerogen volume. Black dots in Track 4 show TOC from lab, brown shading is kerogen volume and red shading is free gas in pore space. Blue dots are core porosity.

Step 6: Identify Gas Intervals

Gas is typically identified by neutron density cross over and  may be masked by the presence of shale or kerogen. Depth plots of shale corrected neutron and density logs will help, but not in kerogen-rich reservoirs, where the crossover due to gas is cancelled by the kerogen effect.

Compatible porosity scales must be appropriate for the interval being evaluated to avoid false positive or negative indications of gas. Running a limestone matrix over a sandstone interval can result in crossover, not caused by the presence of light hydrocarbons.

Calibrate results to gas log or basin knowledge.

Depth plot with shale corrected density and neutron porosity log curves show crossover (shaded areas in Track 3), indicating "obvious" free gas.







Step 7: Identify Coal, Salt and Anhydrite Intervals

Coal intervals are identified by high neutron and density porosity log readings and usually have fairly low GR reading, but not always. Coal zones may be washed out.

Salt is identified by low GR readings, along with a bulk density reading close to 2000 kg/m3, and a neutron porosity close to zero,
sonic log will read 220 us/m. Potash salts (sylvite, carnallite) have high GR and low density. All slats may be washed out due to solution by water-based drilling mud.

Anhydrite is identified by low GR readings, along with a bulk density reading close to 2980 kg/m3, and a neutron porosity value close to zero.

Specific numerical triggers are used to automate this process in software packages. Porosity is set to zero in non-porous zones captured by these triggers.


Calibrate results to known lithology from sample descriptions. Adjust trigger levels to obtain a better match.

Read More about Lithology Triggers


Step 8: Calculate Total Porosity

Total porosity includes clay bound water (CBW). Kerogen will also look like porosity to conventional logs.

Porosity from the neutron density complex lithology crossplot model is the preferred approach and is relatively independent of grain density changes. Other porosity models may also be used; neutron sonic crossplot (less sensitive to bad bore hole conditions), density only (very sensitive to changes in grain density and bore hole conditions), sonic only  (very sensitive to changes in matrix travel time), neutron only (not recommended, a last resort). NMR total porosity is unaffected by kerogen and is independent of mineralogy. It is a good alternate source of total porosity.

Calibrate results to NMR total porosity or low temperature Dean-Stark core analysis. Adjust parameters or select alternate porosity model to obtain a better match.

Read More about Porosity

Step 9: Calculate Effective (Shale and Kerogen Corrected) Porosity

Effective porosity does not include kerogen effects (in kerogen rich reservoirs) or clay bound water. Shale and kerogen corrected versions of the total porosity models described in the previous section are used to calculate effective porosity.


Calibrate results to NMR effective porosity (3 ms cutoff) or high temperature Dean-Stark core analysis, drives off clay bound water, or conventional helium or Boyle's Law core analysis. Adjust parameters or select alternate porosity model to obtain a better match.

Read More about Porosity

Step 10: Calculate Lithology

The lithology model must match the interval being evaluated, and is dependent on available data. Three mineral models from PE, neutron and density logs, or from sonic density and PE logs are best. Two mineral models from sonic or density logs may also be useful. Multi-mineral models should be used with care.

Mineral analysis from logs is required to reconstruct logs for stimulation design.

Calibrate results to XRD mineralogy assay, after converting lab data from mass fraction to volume fraction. Adjust parameters or select alternate lithology model or alternate mineral mixture to obtain a better match.

Read More about Lithology


XRD data used to calibrate clay, quart/feldspar, and carbonate volumes. Doig / Montney interval displaying elemental capture spectroscopy (ECS) processed mineral volumes, which were used for lithology model calibration

Step 11: Calculate Water

The modified Simandoux equation works well for most situations. It accounts for low resistivity clay content and reduces to the Archie equation when volume of shale equals zero. This model is better behaved in low porosity than most other models dual water models may also work, but may give silly results when volume shale is high or porosity is very low.

The tortuosity, cementation and saturation exponents (a, m and n) are required inputs. In many cases electrical properties must be varied from world averages to get SW to match lab data. Recommended values are:
A = 1.0, M = N = 1.5 to 1.8. Lab measurement of electrical properties is essential.

Water resistivity at reference temperature is required and must be corrected to formation temperature. A deep resistivity log reading and accurate shale and kerogen corrected effective porosity are also required.

Calibrate with core SW or capillary pressure data. Adjust RW, A, M, N to obtain better match. Both core SW or capillary pressure data pose problems in unconventional reservoirs, especially reservoirs with thin porosity laminations. Common sense may have to prevail over “facts”.

Read More about Water Saturation

Step 12: Calculate Permeability Index

The Wylie-Rose equation works well in low porosity reservoirs. Calibration constant can range between 100,000 to 150,000 and beyond. Generally assume that the calculated SW is also the irreducible SW. This assumption may not always be correct.

An exponential equation derived from regression of core permeability against core porosity may also work well. High perm data caused by micro or macro fractures should be eliminated before performing the regression.

Permeability from Wyllie-Rose                        Permeability from Regression

Other permeability models are often used. Any model that can be calibrated to core and uses log derived properties will do the job. Most models match conventional core permeability quite well, but will not match permeability derived from crushed samples using the GRI protocol.

Permeability index from log or core analysis must be corrected to in-situ conditions before use in flow capacity or productivity calculations. Log analysis permeability does not include permeability from micro or macro fractures so flow capacity from logs may not match KH from pressure transient analysis. Log perm is usually considered to be matrix permeability.

Calibrate with core permeability, excluding fractured samples. Adjust parameters to obtain a better match.

Read More about Permeability

Step 13: Net Reservoir and Net Pay

Net pay, pore volume, hydrocarbon pore volume and flow capacity are the final result of most petrophysical well log analyses. These are called mappable properties and lead directly to oil and gas in place calculations.

In many shale gas and some shale oil plays, typical porosity cutoffs for net reservoir are very low, 2 or 3% for those with an optimistic view, 4 or 5% for the pessimistic view.

The water saturation cutoff for net pay is quite variable. Some unconventional reservoirs have very little water in the free porosity so the SW cutoff is not too important. Others have higher apparent water saturation than might be expected for a productive reservoir. However, they do produce, so the SW cutoff must be quite liberal. SW cutoffs between 50 and 80% are common.

Shale volume cutoffs are usually quite liberal for unconventional reservoirs, and are usually set above the 50% mark.

Multiple cutoff sets help assess the sensitivity to arbitrary choices and gives an indication of the risk or variability in OGIP or OOIP.

Calibration is not usually possible until years after the field has begun production. May be possible in flowing wells using flowmeter logs.

Read More about Net Pay and Cutoffs

Step 14: Free Gas or Oil in Place

It is easier to compare zones or wells on the basis of OOIP or OGIP instead of average porosity, net pay, or gross thickness. If area = 640 acres and zone thickness is in feet, then OGIP = Bcf/Section (= Bcf/sq.mile) and OOIP is in barrels per square mile. These units of volume are commonly used to compare zones, wells, or different unconventional plays.

Free gas in place is calculated from the usual volumetric equation:
      Bg =  (Ps * (Tf + KT2)) / (Pf * (Ts + KT2)) * ZF 
      OGIPfree = KV4 * PHIe * (1 - Sw) * THICK *  AREA / Bg 
For oil reservoirs:
      OOIP = KV3 * PHIe * (1 - Sw) * THICK *  AREA / Bo

This step is often done by a reservoir engineer on the evaluation team, based on the petrophysical results developed in Steps 1 through 13.

Read More about Gas and Oil in Place.


Step 15: Adsorbed Gas In Place For Kerogen Rich Reservoirs

TOC is widely used as a guide to the quality of shale gas plays. Some deep hot shale gas plays have little adsorbed gas even though they have moderate TOC content. Using correlations of lab measured TOC and gas content (Gc), we can use log derived TOC values to predict Gc. Gc can then be summed over the interval and converted to adsorbed gas in place, again measured in Bcf/section to make it easy to compare projects.

Adsorbed gas in place
Gc = KG11 * TOC%
       OGIPadsorb = KG6 * Gc * DENS * THICK * AREA


Crossplots of TOC versus Gc for Tight Gas / Shale Gas examples. Note the large variation in Gc versus TOC for different rocks, and that the correlations are not always very strong. These data sets are from core samples; cuttings give much worse correlations. The fact that some best fit lines do not pass through the origin suggests systematic errors in measurement or recovery and preservation techniques, and erroneous lost gas estimates.


This step is often done by a reservoir engineer on the evaluation team.

Read More about Adsorbed Gas

Step 16: Reconstruct Sonic and Density Log Curves

For stimulation design modeling, the logs need to represent a water filled reservoir conditions. Since logs read the invaded zone, light hydrocarbons (light oil or gas) make the density log read too low and the sonic log read too high compared to the water filled case. Rock mechanical properties are calculated based on reconstructed logs derived from the petrophysical analysis. The reconstructed logs eliminate gas effect (if any) and low quality data caused by rough borehole.

Calibrate by comparing Vp/Vs ratio (DTS/DTC ratio) with known values for lithology as computed from petrophysical analysis.

Read More about Log Reconstruction

Using bad sonic data results in erroneous elastic properties















Effect of porosity and gas on Poisson’s Ratio. PR will be too low for frac design purposes unless the water filled case is created by log reconstruction.






Example of log reconstruction in a shaly sand sequence (Dunvegan). The 3 tracks on the left show the measured gamma ray, caliper, density, and compressional sonic. Original density and sonic are shown in black, modeled logs are in colour. Shear sonic is the model result as none was recorded in this well. Computed elastic properties are shown in the right hand tracks. Results from the original unedited curves are shown in black, those after log editing are in colour. Note that the small differences in the modeled logs compared to the original curves propagate into larger differences in the results, especially Poisson's Ratio (PR), Young's Modulus (ED), and total closure stress (TCS).

Step 17: Calculate Dynamic
     Mechanical Properties

The reconstructed density and sonic logs are used to calculate:

  •        Poisson’s ratio
               R = DTS / DTC
               PR = (0.5 * R^2 - 1) / (R^2 - 1)

  •        Shear modulus
               N = KS5 * DENS / (DTS ^ 2)  
  •        Young’s dynamic modulus
               Y = 2 * N * (1 + PR)

  •        Bulk modulus
              Kb = KS5 * DENS * (1 / (DTC^2) - 4/3
                       * (1 / (DTS^2)))

  •        Mullin's brittleness index
             Y1 = ((Yst - 1) / (8 - 1) * 100)
             PR1 = ((PR - 0.40) / (0.15 - 0.40)) * 100
             BI = (Y1 + PR1) / 2 


The equations used to generate these values have been used for many years with well log data as input. The results are usually called dynamic rock properties because the sonic log is an impulse (moderately high frequency) measurement. Dynamic measurements can also be made in the lab using a sparker type device. Static measurements are also made in the lab, using pressure sleeves; the process is considered to be a zero frequency or static measurement. Unfortunately, the dynamic and static results do not agree with each other.

Calibrate to dynamic lab data.

Read More about Dynamic Rock Properties  



Step 18: Compare Mechanical Properties to Other Models
Simple linear relationships may work well in clastic intervals, usually relating the parameter to shale volume and mineralogy. Neural network models may also work with corrected log data. The results from the mechanical properties analysis should be compared to the following graphs, based on the lithology and the compressional sonic log values. Data that falls off trend is probably suspect, suggesting that further log editing or adjustments to the analysis parameters are needed.


Young's Modulus versus compressional travel time (DTS)

Poisson's Ratio versus compressional travel time (DTS)

Sample of a mechanical properties log analysis. Left hand tracks show original and reconstructed logs. All results are shown in that right hand tracks. At far right is the lithology / porosity track for correlation.


Step 19: Estimate Static Mechanical Properties

Static values differ from dynamic values because strain and strain rate are dependent on the measurement method.

  •       dynamic: acoustic wave propagation is a phenomenon of small strain at a large strain rate

  •       static (triaxial): large strain at small strain rate

Rocks appear stiffer in response to an elastic wave, compared to a rock mechanics laboratory (triaxial) test.
The weaker the rock, the larger the difference. This accounts for the difference between dynamic and static Young’s moduli. The difference between dynamic and static Poisson’s ratio is very small, and is generally not considered.

Static mechanical rock properties are needed as input for hydraulic fracture simulation work because
static values more closely represent the strain and strain rate created during hydraulic frac stimulation treatments. Many transforms have been published.

Calibrate to static lab data.    

Read More about Static Rock Properties

Step 20: Calculate Closure Stress

Closure stress is calculated using GOHFER’S Total Stress equation and must be calibrated to local field conditions with a strain or stress correction factor. In tectonically active areas, the closure stress calculated from logs will be too low and will need to be increased. Generally, the strain offset approach is favoured.


Pc   = closure pressure, kPa

ν   = Poisson’s Ratio

Dtv   = true vertical depth, m

γob   = overburden stress gradient, kPa/m

γp     = pore fluid gradient, kPa/m

αv   = vertical Biot’s poroelastic constant

αh   = horizontal Biot’s poroelastic constant

Poff   = pore pressure offset, kPa

ε = regional horizontal strain, microstrains

E   = Young’s Modulus, GPa

σt   = regional horizontal tectonic stress, kPa


Overburden Pressure:
The density log is used to calculate overburden stress. The easiest way to calculate overburden stress is by determining the average bulk density above treatment depth. Bad density data are first eliminated by running a discriminator, using  caliper and density correction logs. With the discriminator applied, the average bulk density is calculated and then used to calculate overburden stress.

The more complicated approach requires integration of the bulk density log. This approach requires a synthetic density log to be created. The synthetic log is then integrated from treatment depth to shallowest log reading.

Pore Pressure:
Field measured data should be used to assign pore pressure. Pore fluid supports part of the total stress. Pore pressure depletion increases net stress and leads to compaction. Pore pressure depletion decreases total (fracture closure) stress.

Biot’s constant:
Barree defines Biot’s poro-elastic constant as the efficiency with which internal pore pressure offsets the externally applied vertical total stress. As Biot decreases, net (intergranular) stress increases and pore pressure variations have less impact on net stress.

Calibrate to mini-frac or field data.
The best way to calibrate closure stress is to review previous fracturing work in nearby wells, or to perform a mini-frac. If possible, this step should be completed by the completion engineer (the person running the hydraulic frac simulation software).

Read More about Closure Stress




The following schematic examples, prepared by Dorian Holgate, illustrate computed mechanical properties and closure stress, and how they change with different reservoir conditions.


Base Case: No stress offset, no strain offset         Regional tectonic stress added --- closure stress increases

Mini-frac closure stress does not match base case    Strain offset added to calibrate to field data

Applying a strain offset can decrease the stress difference between the reservoir and non-reservoir intervals - fracture geometry will be affected compared to base case.


Unconventional shale gas example. Results from the custom calculation sequence match SCAL data very well. Next, results were used as input to reconstruct the density and sonic logs. The reconstructed logs were then used to calculate mechanical rock properties.

Clastic Example with Rough Bore Hole. The reconstructed density and sonic logs were used to calculate mechanical rock properties.

See More Examples

Page Views ---- Since 01 Jan 2015
Copyright 1978 - 2017 E. R. Crain, P.Eng. All Rights Reserved